Maintenance of tumor initiating cells of defined genetic
composition by nucleostemin
The MIT Faculty has made this article openly available. Please share
how this access benefits you. Your story matters.
Citation
Okamato, N. et al. "Maintenance of tumor initiating cells of
defined genetic composition by nucleostemin." Proceedings of
the National Academy of Sciences 108.51 (2011): 20388–20393.
Copyright ©2011 by the National Academy of Sciences
As Published
http://dx.doi.org/10.1073/pnas.1015171108
Publisher
National Academy of Sciences
Version
Final published version
Accessed
Wed May 25 22:00:26 EDT 2016
Citable Link
http://hdl.handle.net/1721.1/71896
Terms of Use
Article is made available in accordance with the publisher's policy
and may be subject to US copyright law. Please refer to the
publisher's site for terms of use.
Detailed Terms
Maintenance of tumor initiating cells of defined
genetic composition by nucleostemin
Naoko Okamotoa,b, Mami Yasukawaa, Christine Nguyenc,d, Vivi Kasima, Yoshiko Maidaa, Richard Possematoc,
Tatsuhiro Shibatae, Keith L. Ligonc, Kiyoko Fukamib, William C. Hahnc,e,1, and Kenkichi Masutomia,f,1
a
Division of Cancer Stem Cell, National Cancer Center Research Institute, 5-1-1 Tsukiji, Chuo-ku, Tokyo 104-0045, Japan; bLaboratory of Genome and Biosignal,
Tokyo University of Pharmacy and Life Science, 1432-1 Horinouchi, Hachioji, Tokyo 192-0392, Japan; cDepartment of Medical Oncology, Dana–Farber Cancer
Institute and Department of Medicine, Brigham and Women’s Hospital and Harvard Medical School, Boston, MA 02115; dBroad Institute, Harvard University
and Massachusetts Institute of Technology, Cambridge, MA 02142; eCancer Genomics Project, Center for Medical Genomics, National Cancer Center Research
Institute, 5-1-1 Tsukiji, Chuo-ku, Tokyo 104-0045, Japan; and fPrecursory Research for Embryonic Science and Technology (PREST), Japan Science and
Technology Agency, 4-1-8 Honcho Kawaguchi, Saitama 332-0012, Japan
Edited by Neal F. Lue, Weill Cornell Medical College, New York, NY, and accepted by the Editorial Board June 7, 2011 (received for review October 12, 2010)
Recent work has identified a subset of cells resident in tumors that
exhibit properties similar to those found in normal stem cells. Such
cells are highly tumorigenic and may be involved in resistance to
treatment. However, the genes that regulate the tumor initiating
cell (TIC) state are unknown. Here, we show that overexpression
of either of the nucleolar GTP-binding proteins nucleostemin (NS)
or GNL3L drives the fraction of genetically defined tumor cells that
exhibit markers and tumorigenic properties of TICs. Specifically,
cells that constitutively express elevated levels of NS or GNL3L
exhibit increased TWIST expression, phosphorylation of STAT3,
expression of genes that induce pluripotent stem cells, and enhanced radioresistance; in addition, they form tumors even when
small numbers of cells are implanted and exhibit an increased
propensity to metastasize. GNL3L/NS forms a complex with the
telomerase catalytic subunit [human telomerase reverse transcriptase (hTERT)] and the SWItch-Sucrose NonFermentable (SWI-SNF)
complex protein brahma-related gene 1 (BRG1), and the expression of each of these components is necessary to facilitate the
cancer stem cell state. Together, these observations define a complex composed of TERT, BRG1, and NS/GNL3L that maintains the
function of TICs.
B
oth embryonic and organ-specific stem cells are characterized by the ability to self-renew and to differentiate into
specialized cell types. Recent work has identified a subset of cells
in tumors that exhibit properties similar to those found in normal
stem cells (1–3). Such tumor initiating cells (TICs) or cancer
stem cells are characterized by the capacity for unlimited selfrenewal and the ability to differentiate into multiple cell types.
Moreover, such cells are highly tumorigenic (4), show resistance
to chemotherapy and/or radiotherapy (5–7), and may contribute
to metastasis (8–11).
TICs are defined operationally. Specifically, such cells exhibit
the ability to form tumors when placed in limiting numbers in
animal hosts (4, 12). Although some cancer stem cells express
particular cell surface receptors, the expression of such markers
is not exclusive to TICs and no universal cancer stem cell
markers have been identified. Indeed, recent evidence suggests
that not all tumors may harbor such TICs (12, 13). Given the
potential role for such cells in tumor initiation, progression, and
response to treatment, defining the molecular alterations that
program cancer stem cells is essential not only to identify such
cells but to understand their contribution to malignant transformation.
The nucleolar GTP-binding protein nucleostemin (NS) and its
closely related family member GNL3L are expressed at high
levels in ES cells (14, 15), and NS has been proposed as a marker
for TICs in highly aggressive brain tumors (16). NS has been
reported to regulate cell proliferation through a direct interaction with p53 (14). Specifically, recent studies have shown
that expression of NS and/or GNL3L delays the onset of cellular
senescence by negatively regulating telomeric repeat-binding
factor 1 (TRF1) stability (17) and that depletion of NS causes G1
20388–20393 | PNAS | December 20, 2011 | vol. 108 | no. 51
arrest in a p53-dependent manner (18). However, other studies
suggest that NS may also contribute to stem cell function independent of p53, because blastocysts derived from NS null mice
failed to enter S phase even in the absence of p53 (19).
We hypothesized that NS may contribute directly to formation
of cancer stem cells. Here, we show that the expression of NS/
GNL3L increases the fraction of tumorigenic human cells that
exhibit TIC properties.
Results
NS and GNL3L Regulate TIC Behavior. To examine whether endog-
enous NS was essential for the behavior of established TICs, we
evaluated the consequences of suppressing NS in well-characterized TIC lines derived from glioblastomas (GBMs; 0308, BT145,
and BT112) (20). To determine whether suppression of NS in these
GBM TICs affected clonogenic neurosphere formation, a phenotype tightly correlated with tumorigenicity (21), cells were infected
with either a control (short hairpin GFP) or three distinct NSspecific shRNAs [short hairpin NS (shNS) 1–3; Fig. 1]. We performed single-cell neurosphere formation assays and found that
cells expressing the control shRNA formed neurospheres comparable to unmanipulated GBM TICs. In contrast, 0308 cells expressing NS-specific shRNAs displayed altered sphere-forming
activity that corresponded to the degree of NS suppression. Specifically, cells expressing NS-specific shRNAs that produced the
highest degree of NS suppression (shNS1 and shNS3) exhibited
a significant decrease in average sphere size (P < 0.02; Fig. 1A,
Upper Left). These cells also displayed decreased levels of neural
stem cell markers CD133 and SOX2 (Fig. 1A, Lower Left). To
eliminate the possibility that the observed effects were attributable
to off-target effects of RNAi, we created a shRNA specific for the
NS 3′UTR (shNS4) and showed that expression of this shRNA
induced similar effects on the proliferation of MCF7 cells as the
other NS-specific shRNAs and could be rescued by the expression
of ectopic NS (Fig. S1 A–C). Moreover, suppression of NS in two
other TIC lines, BT145 and BT112 (Fig. 1 B and C), also resulted in
This paper results from the Arthur M. Sackler Colloquium of the National Academy of
Sciences, “Telomerase and Retrotransposons: Reverse Transcriptases That Shaped Genomes,” held September 29–30, 2010, at the Arnold and Mabel Beckman Center of the
National Academies of Sciences and Engineering in Irvine, CA. The complete program and
audio files of most presentations are available on the NAS Web site at www.nasonline.
org/telomerase_and_retrotransposons.
Author contributions: N.O., C.N., Y.M., R.P., T.S., K.F., W.C.H., and K.M. designed research;
N.O., M.Y., C.N., V.K., Y.M., and R.P. performed research; N.O., M.Y., C.N., R.P., T.S., K.L.L.,
K.F., W.C.H., and K.M. contributed new reagents/analytic tools; N.O., M.Y., C.N., V.K., Y.M.,
R.P., T.S., K.L.L., K.F., W.C.H., and K.M. analyzed data; and N.O., C.N., W.C.H., and K.M.
wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission. N.F.L. is a guest editor invited by the Editorial
Board.
1
To whom correspondence may be addressed. E-mail: william_hahn@dfci.harvard.edu or
kmasutom@ncc.go.jp.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.
1073/pnas.1015171108/-/DCSupplemental.
www.pnas.org/cgi/doi/10.1073/pnas.1015171108
SACKLER SPECIAL
FEATURE
human cancer lines (HeLa or MCF7), we expressed NS or
GNL3L at levels found in GBM-derived neurospheres and analyzed several phenotypes associated with TICs. We found that
expression of either NS or GNL3L conferred the ability to form
statistically larger and more numerous anchorage-independent
colonies (Fig. 2 A and B) and induced enhanced tumorigenicity
(Fig. 2C). Specifically, we observed tumor formation after the s.c.
inoculation of a small number of cells overexpressing NS or
GNL3L (≤500 cells), whereas cells expressing a control vector
failed to form tumors. TICs have also been reported to exhibit
relative resistance to ionizing radiation (5). When we irradiated
BJ-hTERT, MCF7, or HeLa cells expressing a control vector,
NS, or GNL3L, we found that cells expressing NS or GNL3L
were more resistant to γ-irradiation (Fig. S2). These observations
suggest that expression of NS or GNL3L alters the fraction of
TICs present in both established and engineered cancer cell lines.
NS or GNL3L Expression Induces Markers and Pathways Associated
with TICs. Several markers have been described to be expressed
by TICs. When we analyzed cells overexpressing NS or GNL3L,
we found increased expression of CD44 and CD133, markers
associated with TICs (4, 24) (Fig. 3 A–C). Indeed, we found that
the expression of NS or GNL3L in HeLa cells and BJ-hTERT
cells induced a two- to fourfold increase in the percentage of
cells expressing high levels of CD44 compared with cells expressing a control vector when we assessed either total CD44
levels (Fig. 3A) or CD44 expression in individual cells (Fig. 3 B
and C). In addition, we confirmed that NS and GNL3L are
a decrease in sphere size and expression of the neural stem cell
marker CD133 proportional to the degree of NS suppression. Together, these observations indicate that suppression of NS in 0308,
BT145, and BT112 cells decreases the ability of these cells to form
clonogenic neurospheres, and likely their tumor initiating capacity.
Although tumorigenic, human cancer cell lines derived from
tumors or engineered by the expression of specific genetic elements (22) contain only a small fraction of cells that exhibit TIC
behavior (4, 10, 23). To determine the effects of NS or GNL3L
on the tumorigenicity of either engineered tumorigenic fibroblast
[BJELR (full name: BJEHEcR-LT-Ras-ST)] and kidney epithelial (HA1ER) cells expressing the SV40 large and small T
antigens, hTERT, and oncogenic H-RAS (22) or two established
Okamoto et al.
CELL BIOLOGY
Fig. 1. NS is required for TIC function. (A, Upper Left) Suppression of NS in
0308 cells results in decreased neurosphere size. The 0308 cells expressing
control or NS-specific shRNAs were grown under neurosphere promoting
conditions. Triplicate images of each cell population for three independent
experiments were analyzed for average neurosphere size per area (μm2).
Error bars indicate SD for three independent experiments. (Lower Left)
Control (short hairpin GFP) and NS-specific shRNAs (shNS1–3) were introduced into 0308 cells, and endogenous NS suppression was measured
by immunoblotting. IB, immunoblot; shGFP, short hairpin GFP. The expression of CD133 and SOX2 was also assessed; β-actin is shown as a loading
control. (Right) Representative images of GBM-derived 0308 TICs containing
a control or NS-specific shRNAs. (Scale bar = 250 μm.) As described in A
performed on BT145 cells (B) or BT112 cells (C). (Scale bar = 1 mm in representative images.)
Fig. 2. Expression of NS or GNL3L induces TICs. (A) Anchorage-independent
growth. For each cell line, 100 cells were plated and colony numbers were
counted after 1 mo. The mean ± SD for three independent experiments is
shown. *P < 0.05. (B) Representative micrographs demonstrate colony sizes
(magnification: 100×) of HeLa cells expressing NS, GNL3L, or a control vector
(control). (Scale bar = 100 μm.) (C) For tumorigenicity assays, the indicated
numbers of cells were injected s.c. into immunodeficient mice and are
reported as the number of tumors formed/number of injection sites.
PNAS | December 20, 2011 | vol. 108 | no. 51 | 20389
significantly overexpressed in the subpopulation of several cancer cell lines previously shown (4, 23, 25) to be enriched in cells
that express gene signatures related to TICs (CD44 high fraction
of HeLa and CD44 high/CD24 low fraction of MDA-MB-157,
MDA-MB-231, and MDA-MB-436 cells; Fig. S3 A and B).
Cell surface markers are useful tools to identify cells that exhibit cancer stem cell activity but may not unequivocally mark
this cell population. We also investigated whether NS expression
alters pathways implicated in the maintenance of cancer stem
cells. Specifically, we determined whether expression of NS or
GNL3L affected WNT/β-catenin signaling and found that expression of NS or GNL3L correlated with increased nuclear localization of the active nonphosphorylated β-catenin (Fig. 3 D
and E) and induced statistically significant increased β-catenin
activity as measured by reporter assay (26, 27) (Fig. 3F). Expression of NS or GNL3L also induced higher steady-state levels
of c-MYC, a direct β-catenin target gene (Fig. 3G). The introduction of OCT3/4, SOX2, c-MYC, and KLF4 into normal
human and murine cells suffices to reprogram such cells into ES
cells (28, 29). When we analyzed the expression of these genes in
cells expressing either a control vector, NS, or GNL3L, we found
that NS or GNL3L up-regulated c-MYC, OCT3/4, and KLF4
(Fig. 3G and Fig. S4). These observations confirmed that the
expression of NS or GNL3L induces the expression of genes
associated with ES cells and cancer stem cells.
NS or GNL3L Expression Activates the Epithelial Mesenchymal Transition and Induces Metastasis. In addition to WNT/β-catenin sig-
naling, several other pathways have been implicated in the
maintenance of the cancer stem cell state. Specifically, recent
evidence indicates that increased STAT3 signaling regulates the
expression of the master regulator TWIST to induce epithelial
mesenchymal transition (EMT) and metastasis (30). In consonance with these observations, we found that BJ-hTERT, MCF7,
and HeLa cells expressing NS or GNL3L exhibited increased
expression of the tyrosine phosphorylated form of STAT3 (Tyr705;
Fig. 4A) and higher levels of TWIST, SNAIL, and vimentin (Fig.
4B). TGF-β signaling has been implicated in the regulation of
both CD44 expression and TWIST (10). To investigate whether
TGF-β signaling was required for EMT in cells expressing NS or
GNL3L, we treated HeLa cells expressing NS with two different
TGF-β inhibitors (LY364947 and SD-208) and found that the
expression of CD44 was decreased in HeLa cells expressing NS in
a dose-dependent manner (Fig. 4E). These observations suggest
that the expression of NS or GNL3L induces an EMT.
In addition, we found that tumorigenic cells expressing either
NS or GNL3L exhibited increased capacity for metastasis as
measured by the number of foci formed in the lungs of mice
after tail vein injection of control, NS-expressing, or GNL3Lexpressing cells (Fig. 4 C and D). Although this observed increase in the number of lung metastases in cells expressing NS
may be attributable, in part, to the increased tumorigenicity of
these cells, these observations show that NS-expressing cells
exhibit several phenotypes associated with TICs (10, 11).
NS/GNL3L, BRG1, and TERT Form a Complex Necessary for TIC
Function. Because TERT has been implicated in the mainte-
nance of stem cells by both telomere-dependent (31) and telomereindependent mechanisms (32–35) and had previously been
reported to interact with GNL3L (36), we determined whether
the interaction of NS/GNL3L with hTERT was involved in the
induction of TICs. Specifically, we isolated hTERT immune
complexes and found that endogenous NS interacts with endogenous hTERT in both HeLa and 293T cells (Fig. 5A). To
characterize the interaction of hTERT with NS or GNL3L, we
used a series of TERT truncation mutants and found that the
amino-terminal end of hTERT (1–531) was necessary for hTERT
to interact with both NS and GNL3L (Fig. S5A). We also found
that the conserved consensus motifs G5 and G4 present in both
NS and GNL3L (37–40) were required for the interaction with
hTERT (Figs. S5 B and C, and S6). These observations identify
20390 | www.pnas.org/cgi/doi/10.1073/pnas.1015171108
Fig. 3. Effects of NS or GNL3L on cancer stem cell markers. Effects of overexpressing NS or GNL3L on CD44 protein expression as assessed by immunoblotting (A) or flow cytometry (B and C). Cells were stained with FITCconjugated anti-CD44 (Leu-44) antibody. IB, immunoblot. (B) Fractions of
HeLa cells expressing high levels of CD44 were 6.8% (control vector), 24.0%
(NS), and 14.1% (GNL3L). (C) Fractions of BJ-hTERT cells expressing high levels
of CD44 were 8.3% (control vector), 27.3% (NS), and 17.5% (GNL3L). (D)
Percentage of cells that harbor nuclear activated β-catenin in cells expressing
a control vector, NS, or GNL3L is shown. *P < 0.05. (E) Effects of overexpressing NS or GNL3L on β-catenin function. Representative immunofluorescence images are shown. HeLa cells (Upper) and HeLa cells expressing NS
(Lower) were stained with an antibody that recognizes unphosphorylated
(active) β-catenin and visualized with Alexa Fluor 488 (Invitrogen)-conjugated
anti-mouse IgG (green, magnification: 600×); DAPI (blue) indicates DNA.
(Scale bar = 10 μm.) (F) TOPFLASH-Luc luciferase reporter activity. A renilla
luciferase expression plasmid, pRL-SV40, was used as an internal control for
transfection efficiency. *P < 0.05. (G) Effects of overexpressing NS or GNL3L
on the expression of c-MYC, OCT3/4, KLF4, and SOX2.
Okamoto et al.
a specific conserved region in NS and GNL3L responsible for the
interaction with hTERT.
As recently described (35), we confirmed that hTERT binds
both overexpressed and endogenous BRG1 (Fig. 5 B and C). To
determine whether hTERT, BRG1, and NS/GNL3L are found in
Fig. 5. NS forms a complex with hTERT and BRG1. (A) hTERT
interacts with endogenous NS. hTERT complexes from 293T or
HeLa cells were purified with an anti-hTERT antibody (Rockland), and associated proteins were subjected to SDS/PAGE and
immunoblotting with an anti-NS antibody. IB, immunoblot. (B)
hTERT interacts with endogenous BRG1. hTERT immune complexes from 293T or HeLa cells were subjected to SDS/PAGE,
followed by immunoblotting with an anti-BRG1 antibody. (C)
BRG1 interacts with endogenous NS. BRG1 complexes from
293T cells were purified by immunoprecipitation with an antiBRG1 antibody and subjected to SDS/PAGE, followed by immunoblotting with an anti-NS antibody. Schema (D) and
results (E) of the sequential immunoprecipitation using cells
expressing FLAG-tagged NS or GNL3L and hTERT-HA. NS or
GNL3L immune complexes were isolated and then eluted with
the FLAG M2 peptide. hTERT immune complexes were then
purified, and endogenous BRG1 was detected by immunoblotting. Schema (F) and results (G) of the sequential immunoprecipitation using cells stably expressing TAP-hTERT. hTERT
immune complexes were isolated and eluted with the FLAG
M2 peptide. Endogenous BRG1 immune complexes were isolated, and endogenous NS was analyzed by immunoblotting.
Okamoto et al.
PNAS | December 20, 2011 | vol. 108 | no. 51 | 20391
SACKLER SPECIAL
FEATURE
CELL BIOLOGY
Fig. 4. Effects of NS and GNL3L on EMT and metastasis. (A) Effects of
overexpression of NS or GNL3L on STAT3 and phospho-STAT3 (Tyr-705). IB,
immunoblot. (B) Effects of overexpression of NS or GNL3L on TWIST, SNAIL,
and vimentin expression. (C and D) Effects of overexpressing NS or GNL3L on
metastasis. (C) Bar graph shows the number of metastatic foci found after
tail vein injection. The mean ± SD for three independent experiments is
shown. *P < 0.05. (D) Representative H&E images are shown (magnification:
100×). (Scale bar = 50 μm.) (E) Effects of TGF-β inhibitors on CD44 expression
induced by NS/GNL3L. The indicated cells were treated with LY364947 or SD208 for 24 h, followed by immunoblotting.
the same complex, we performed two types of sequential immunoprecipitation experiments. First, we isolated FLAG–NS or
FLAG–GNL3L immune complexes and confirmed that hTERTHA was present. We then eluted these complexes with the M2
(FLAG) peptide, isolated hTERT immune complexes, and confirmed that endogenous BRG1 was present in these complexes
(Fig. 5 D and E). Second, we isolated hTERT immune complexes
and confirmed that BRG1 was present. We then eluted these
complexes with the M2 (FLAG) peptide, isolated BRG1 immune
complexes, and confirmed that endogenous NS was present (Fig.
5 F and G). These observations suggest that hTERT, BRG1, and
NS or GNL3L are present in the same complex.
To determine whether the interaction of hTERT with NS/
GNL3L required the telomere elongation activities of TERT, we
suppressed the expression of the telomerase RNA subunit
hTERC and found that hTERT continued to associate with NS,
indicating that this interaction occurs independent of telomerase
activity (Fig. S7A). Moreover, overexpression or suppression of
NS or GNL3L did not affect telomerase activity or telomere
length as assessed by telomere repeat amplification protocol
assays or telomere restriction fragment Southern blotting (Fig.
S7 B–D). Together, these observations suggest that complexes
containing hTERT, BRG1, and NS or GNL3L do not contribute
directly to telomere maintenance but, instead, drive TIC formation by regulating the activity of BRG1.
To confirm that the complex composed of hTERT, BRG1,
and NS/GNL3L was involved in regulating TIC phenotypes, we
assessed whether each of these components was necessary for
NS-induced increase in CD44 levels. When we suppressed
hTERT expression using hTERT-specific siRNAs, we failed to
observe an increase in CD44 expression after expressing NS or
GNL3L (Fig. 6A). Moreover, BRG1 suppression reversed the
increase in CD44 expression induced by NS or GNL3L expression (Fig. 6B). In contrast, when we suppressed hTERC expression using hTERC-specific shRNAs, we found that the expression
of hTERC was not required for the increase in CD44 expression
induced by NS or GNL3L overexpression (Fig. S7E). Furthermore, when we suppressed hTERT or BRG1 expression, we failed
to observe an increase in TWIST expression after expressing NS
or GNL3L (Fig. 6 C and D). Suppression of hTERT or BRG1 in
cells expressing NS or GNL3L, which are markedly enriched in
TICs, ablated the ability of HeLa cells expressing NS or GNL3L
to form tumors (Fig. 6 E and F). Together, these observations
Fig. 6. Complex of hTERT, BRG1, and NS/GNL3L is necessary for NS-induced
TIC formation. (A) hTERT is required for NS/GNL3L-induced increase in CD44
expression. HeLa-NS cells or HeLa-GNL3L cells expressing either a control
siRNA or two independent hTERT-specific siRNAs were immunoblotted for
CD44. IB, immunoblot. (B) BRG1 is required for NS/GNL3L-induced increase in
CD44 expression. CD44 expression was assessed by immunoblotting in HeLaNS cells or HeLa-GNL3L cells stably expressing either a control vector or two
independent BRG1-specific shRNAs. (C) hTERT is required for NS/GNL3Linduced increase in TWIST expression. HeLa-NS cells or HeLa-GNL3L cells
expressing either a control siRNA or two independent hTERT-specific siRNAs
were immunoblotted for TWIST. (D) BRG1 is required for NS/GNL3L-induced
increase in TWIST expression. TWIST expression was assessed by immunoblotting in HeLa-NS cells or HeLa-GNL3L cells stably expressing either a control vector or two independent BRG1-specific shRNAs. (E) hTERT is required
for tumor formation by HeLa cells. The number of tumors formed/number of
injection sites in HeLa-NS cells or HeLa-GNL3L cells expressing either a control siRNA or two independent hTERT-specific siRNAs is shown. For E and F,
5 × 104 cells were injected per mouse. NC, negative control. (F) BRG1 is required for tumor formation by HeLa cells expressing NS or GNL3L. HeLa-NS
cells or HeLa-GNL3L cells expressing either a control vector or two independent BRG1-specific shRNAs are shown.
indicate that the expression of hTERT, NS/GNL3L, and BRG1 is
each required to induce a TIC phenotype.
Discussion
Here, we show that expression of NS or GNL3L in genetically
engineered human cancer cells confers a TIC phenotype. Tumorigenic human cells of defined genetic constitution expressing
NS or GNL3L exhibited the full range of phenotypes associated
with cancer stem cells, including the expression of cell surface
markers such as CD44 and CD133, genes that induce induced
pluripotant stem (iPS) cells, the ability to form tumors when
implanted at limiting numbers, activation of an EMT program,
and increased metastatic potential. Because genetically engineered human cancer cells (BJELR and HA1ER) and even
established human cancer lines (HeLa and MCF7) have both
TIC and non-TIC populations (Fig. S3), expressing NS or
GNL3L enhances the TIC fraction.
20392 | www.pnas.org/cgi/doi/10.1073/pnas.1015171108
NS was initially found to be expressed at high levels in highly
proliferative multipotential cells (14, 19) as well as a subset of
aggressive brain tumors (16). We found that NS and GNL3L
contribute directly to the maintenance of the TIC phenotype.
Prior work demonstrated that NS physically interacts with p53 and
that this interaction regulates cell proliferation (14). However, it
is clear that NS also contributes to stem cell function independent of p53 (19). Although we have confirmed that depletion of
NS is associated with cell cycle arrest accompanied by p53 upregulation (18), the tumorigenic cells in which we expressed NS or
GNL3L lack p53 function, suggesting that NS contributes to the
generation of TIC independent of its regulation of p53.
TERT was recently shown to interact with GNL3L in a manner
that regulated telomere length (36). We have confirmed that
GNL3L and NS interact with TERT through a highly conserved
region present in both of these nucleolar GTP binding proteins.
However, in this setting, we failed to find evidence that the interaction of NS or GNL3L affected telomere length or telomerase activity. Instead, we found that NS or GNL3L forms
a ternary complex with TERT and BRG1 in a manner that does
not depend on hTERC. However, it remains possible that TERT–
GNL3L complexes regulate telomere dynamics in other cells or
conditions. These observations suggest that the NS/GNL3L–
TERT–BRG1 complex operates in a telomere-independent
manner to drive transcriptional programs essential to the maintenance of the TIC state. Thus, in addition to facilitating immortalization, hTERT may contribute to tumorigenicity through
the NS/GNL3L–TERT–BRG1 complex by influencing the balance of cancer cells that exhibit TIC phenotypes.
Recent work indicates that TERT regulates stem cell homeostasis independent of its function at telomeres (32) and that
TERT, together with BRG1, modulates stem cell homeostasis
by modulating the WNT/β-catenin signaling pathway (35). We
confirmed that the NS/GNL3L–TERT–BRG1 complex is essential for the maintenance of TICs. Further work will be necessary to determine whether this complex regulates normal stem
cells by a similar mechanism.
In prior work, we identified a limited number of introduced
genes that sufficed to convert normal human cells into tumorigenic cells (22). We and others have subsequently found specific
combinations of genes mutating in spontaneously arising cancers
that permitted the creation of tumorigenic human cells of a defined genetic composition (41–43). These experimental models
have proven useful to decipher the cooperation between mutated
genes and the discovery of new cancer genes (44, 45). However,
like most established cancer cell lines, such cells formed tumors
only after the implantation of a large number of cells, suggesting
that additional perturbations were necessary to drive the formation of TICs. The observation that expression of NS or
GNL3L in this setting leads to cells that exhibit all the phenotypes associated with cancer stem cells establishes an experimental model to define the pathways required to program the
cancer stem cell state. Moreover, because the acquisition of stem
cell characteristics correlates with treatment resistance and metastatic disease, these experimental models may provide the
means to identify new targets to inhibit these phenotypes associated with lethal disease.
Materials and Methods
Cell Culture and Stable Expression of FLAG-NS/GNL3L and hTERT. The human
cell lines 293T, MCF7, HeLa-S, HA1ER (22), and HeLa were maintained in
DMEM supplemented with 10% (vol/vol) heat-inactivated FBS. BJ fibroblasts
and BJELR cells were cultured as described (22). Amphotropic retroviruses
were created as described (22) using the expression vector pWZL-neo,
pBABE-Hygro, pWZL-neo-FLAG-NS, pWZL-neo-FLAG-GNL3L, or pBABEHygro-hTERT. After infection, cells were selected with neomycin (G418; 2
mg/mL) for 7 d or with hygromycin (50 mg/mL) for 3 d.
Stable Expression of shRNA. The pLKO.1-puro vector was used to express
shRNAs targeting NS, GNL3L, hTERC, BRG1, and GFP. These vectors were
used to make amphotropic lentiviruses, and polyclonal cell populations were
purified by selection with puromycin (2 mg/mL).
Okamoto et al.
Anchorage-Independent Growth and Tumorigenicity Assays. Growth in soft
agar was performed as described (22) and scored at 4 wk. For tumor
experiments, cells were mixed with BD Matrigel Matrix (BD Bioscience) at 4 °C
and injected s.c. in BALB/c-nu/nu mice.
Metastasis Assay. A total of 5 × 106 cells were injected into the tail vein of
BALB/c-nu/nu mice. After 4 wk, lungs were dissected to evaluate tissue
morphology and to detect metastases.
Neurosphere Formation Assay. GBM-derived 0308 tumor stem cells were a gift
from H. Fine (US National Institutes of Health, Bethesda, MD) and were
cultured in neurobasal media (NBE; Invitrogen), 0.5× each N2 and B27 supplements (Invitrogen), and 50 ng/mL each human recombinant basic FGF and
EGF (R&D Systems). BT145 and BT112 cells were cultured in NeuroCult NS-A
Basal Medium (StemCell Technologies) supplemented with NeuroCult NS-A
Proliferation Supplement (StemCell Technologies) and 20 ng/mL human
recombinant bFGF and EGF. For shRNA experiments, cells were infected with
viral supernatant diluted 1:5, spun at 930 × g for 30 min, and incubated at
37 °C for 1.5 h before changing to fresh media. Cells were selected with 0.4
mg/mL puromycin 24 h after infection. At 72 h after infection, the cells were
ACKNOWLEDGMENTS. We thank A. Miyajima, S. Saito, S. Inanobe,
S. Takahashi, T. Ochiya, and R. Takahashi for their technical assistance. We
thank H. Fine for the gift of 308 cells. This work was supported, in part, by
Grant-in-Aid for Young Scientists (A) 21689012 (to K.M.) from the Japanese
Ministry of Education, Culture, Sports, Science, and Technology; by ThirdTerm Comprehensive Control Research for Cancer (K.M.) from the Japanese
Ministry of Health, Labor, and Welfare; by the NOVARTIS Foundation
(Japan) for the Promotion of Science (K.M.); by the Funding Program for
Next Generation World-Leading Researchers (K.M.); and by Grant R01
AG23145 from the US National Institutes of Health (to W.C.H.). N.O. was
supported by a Research Fellow of the Japan Society for the Promotion
of Science.
1. Reya T, Morrison SJ, Clarke MF, Weissman IL (2001) Stem cells, cancer, and cancer stem
cells. Nature 414:105–111.
2. Dalerba P, Cho RW, Clarke MF (2007) Cancer stem cells: Models and concepts. Annu
Rev Med 58:267–284.
3. Dick JE (2009) Looking ahead in cancer stem cell research. Nat Biotechnol 27:44–46.
4. Al-Hajj M, Wicha MS, Benito-Hernandez A, Morrison SJ, Clarke MF (2003) Prospective
identification of tumorigenic breast cancer cells. Proc Natl Acad Sci USA 100:
3983–3988.
5. Bao S, et al. (2006) Glioma stem cells promote radioresistance by preferential
activation of the DNA damage response. Nature 444:756–760.
6. Rich JN, Bao S (2007) Chemotherapy and cancer stem cells. Cell Stem Cell 1:353–355.
7. Todaro M, et al. (2007) Colon cancer stem cells dictate tumor growth and resist cell
death by production of interleukin-4. Cell Stem Cell 1:389–402.
8. Dalerba P, Clarke MF (2007) Cancer stem cells and tumor metastasis: First steps into
uncharted territory. Cell Stem Cell 1:241–242.
9. Hermann PC, et al. (2007) Distinct populations of cancer stem cells determine tumor
growth and metastatic activity in human pancreatic cancer. Cell Stem Cell 1:313–323.
10. Mani SA, et al. (2008) The epithelial-mesenchymal transition generates cells with
properties of stem cells. Cell 133:704–715.
11. Nagahama Y, et al. (2010) PSF1, a DNA replication factor expressed widely in stem
and progenitor cells, drives tumorigenic and metastatic properties. Cancer Res 70:
1215–1224.
12. Quintana E, et al. (2008) Efficient tumour formation by single human melanoma cells.
Nature 456:593–598.
13. Shackleton M, Quintana E, Fearon ER, Morrison SJ (2009) Heterogeneity in cancer:
Cancer stem cells versus clonal evolution. Cell 138:822–829.
14. Tsai RY, McKay RD (2002) A nucleolar mechanism controlling cell proliferation in stem
cells and cancer cells. Genes Dev 16:2991–3003.
15. Baddoo M, et al. (2003) Characterization of mesenchymal stem cells isolated from
murine bone marrow by negative selection. J Cell Biochem 89:1235–1249.
16. Tamase A, et al. (2009) Identification of tumor-initiating cells in a highly aggressive
brain tumor using promoter activity of nucleostemin. Proc Natl Acad Sci USA 106:
17163–17168.
17. Zhu Q, Yasumoto H, Tsai RY (2006) Nucleostemin delays cellular senescence and
negatively regulates TRF1 protein stability. Mol Cell Biol 26:9279–9290.
18. Ma H, Pederson T (2007) Depletion of the nucleolar protein nucleostemin causes G1
cell cycle arrest via the p53 pathway. Mol Biol Cell 18:2630–2635.
19. Beekman C, et al. (2006) Evolutionarily conserved role of nucleostemin: Controlling
proliferation of stem/progenitor cells during early vertebrate development. Mol Cell
Biol 26:9291–9301.
20. Lee J, et al. (2006) Tumor stem cells derived from glioblastomas cultured in bFGF and
EGF more closely mirror the phenotype and genotype of primary tumors than do
serum-cultured cell lines. Cancer Cell 9:391–403.
21. Vescovi AL, Galli R, Reynolds BA (2006) Brain tumour stem cells. Nat Rev Cancer 6:
425–436.
22. Hahn WC, et al. (1999) Creation of human tumour cells with defined genetic
elements. Nature 400:464–468.
23. Shipitsin M, et al. (2007) Molecular definition of breast tumor heterogeneity. Cancer
Cell 11:259–273.
24. Singh SK, et al. (2003) Identification of a cancer stem cell in human brain tumors.
Cancer Res 63:5821–5828.
25. Ben-Porath I, et al. (2008) An embryonic stem cell-like gene expression signature in
poorly differentiated aggressive human tumors. Nat Genet 40:499–507.
26. Miller JR, Hocking AM, Brown JD, Moon RT (1999) Mechanism and function of signal
transduction by the Wnt/beta-catenin and Wnt/Ca2+ pathways. Oncogene 18:7860–7872.
27. Firestein R, et al. (2008) CDK8 is a colorectal cancer oncogene that regulates betacatenin activity. Nature 455:547–551.
28. Takahashi K, Yamanaka S (2006) Induction of pluripotent stem cells from mouse
embryonic and adult fibroblast cultures by defined factors. Cell 126:663–676.
29. Takahashi K, et al. (2007) Induction of pluripotent stem cells from adult human
fibroblasts by defined factors. Cell 131:861–872.
30. Cheng GZ, et al. (2008) Twist is transcriptionally induced by activation of STAT3 and
mediates STAT3 oncogenic function. J Biol Chem 283:14665–14673.
31. Flores I, Cayuela ML, Blasco MA (2005) Effects of telomerase and telomere length on
epidermal stem cell behavior. Science 309:1253–1256.
32. Sarin KY, et al. (2005) Conditional telomerase induction causes proliferation of hair
follicle stem cells. Nature 436:1048–1052.
33. Choi J, et al. (2008) TERT promotes epithelial proliferation through transcriptional
control of a Myc- and Wnt-related developmental program. PLoS Genet 4:e10.
34. Imamura S, et al. (2008) A non-canonical function of zebrafish telomerase reverse
transcriptase is required for developmental hematopoiesis. PLoS ONE 3:e3364.
35. Park JI, et al. (2009) Telomerase modulates Wnt signalling by association with target
gene chromatin. Nature 460:66–72.
36. Fu D, Collins K (2007) Purification of human telomerase complexes identifies factors
involved in telomerase biogenesis and telomere length regulation. Mol Cell 28:
773–785.
37. Daigle DM, et al. (2002) YjeQ, an essential, conserved, uncharacterized protein from
Escherichia coli, is an unusual GTPase with circularly permuted G-motifs and marked
burst kinetics. Biochemistry 41:11109–11117.
38. Leipe DD, Wolf YI, Koonin EV, Aravind L (2002) Classification and evolution of P-loop
GTPases and related ATPases. J Mol Biol 317:41–72.
39. Meng L, Yasumoto H, Tsai RY (2006) Multiple controls regulate nucleostemin
partitioning between nucleolus and nucleoplasm. J Cell Sci 119:5124–5136.
40. Yasumoto H, Meng L, Lin T, Zhu Q, Tsai RY (2007) GNL3L inhibits activity of estrogenrelated receptor gamma by competing for coactivator binding. J Cell Sci 120:
2532–2543.
41. Zhao JJ, et al. (2003) Human mammary epithelial cell transformation through the
activation of phosphatidylinositol 3-kinase. Cancer Cell 3:483–495.
42. Boehm JS, Hession MT, Bulmer SE, Hahn WC (2005) Transformation of human and
murine fibroblasts without viral oncoproteins. Mol Cell Biol 25:6464–6474.
43. Zhao JJ, et al. (2005) The oncogenic properties of mutant p110alpha and p110beta
phosphatidylinositol 3-kinases in human mammary epithelial cells. Proc Natl Acad Sci
USA 102:18443–18448.
44. Kolfschoten IG, et al. (2005) A genetic screen identifies PITX1 as a suppressor of RAS
activity and tumorigenicity. Cell 121:849–858.
45. Boehm JS, et al. (2007) Integrative genomic approaches identify IKBKE as a breast
cancer oncogene. Cell 129:1065–1079.
Okamoto et al.
PNAS | December 20, 2011 | vol. 108 | no. 51 | 20393
SACKLER SPECIAL
FEATURE
trypsinized and plated to assay for neurosphere formation; after 72 h, imaging with a light microscope was performed and cell lysates were made for
immunoblot analysis. To determine average sphere size, images were analyzed using ImageJ software (National Institutes of Health).
CELL BIOLOGY
Immunoblotting and Immunoprecipitation. Cells were lysed using a Nonidet
P-40/SDS–based lysis buffer. Details concerning the conditions and antibodies used are found in SI Materials and Methods.